On 01/19/2018 10:55 AM, Ján Tomko wrote:
> On Fri, Jan 19, 2018 at 10:02:45AM +0100, Michal Privoznik wrote:
>> After aeda1b8c56dc5 we tried to stop reporting I/O errors on
>> expected monitor HUP. We've achieved that by not setting
>> mon->lastError once we've received SHUTDOWN event. However, thi
On Fri, Jan 19, 2018 at 10:02:45AM +0100, Michal Privoznik wrote:
After aeda1b8c56dc5 we tried to stop reporting I/O errors on
expected monitor HUP. We've achieved that by not setting
mon->lastError once we've received SHUTDOWN event. However, this
makes us to deadlock in case there's thread that
After aeda1b8c56dc5 we tried to stop reporting I/O errors on
expected monitor HUP. We've achieved that by not setting
mon->lastError once we've received SHUTDOWN event. However, this
makes us to deadlock in case there's thread that enters the
monitor after the event is received. The problem is, we'
